FAQ Search Today's Posts Mark Forums Read
» Video Reviews

» Linux Archive

Linux-archive is a website aiming to archive linux email lists and to make them easily accessible for linux users/developers.


» Sponsor

» Partners

» Sponsor


 
 
LinkBack Thread Tools
 
Old 06-13-2008, 06:37 AM
Roy Marples
 
Default

On Friday 13 June 2008 02:13:19 David Leverton wrote:
> The pkgcore was (or should have been) highly obvious to anyone who had
> so much glanced at the offending code.

Good behaviour
Hey - I found this bug in your code.
Here's a patch!

Bad behaviour
Hey guys - stop using Foo as it has a highly obvious bug which *would* have
been caught in automated testing. This is a clear demonstration that Foo
sucks and you should use Bar!

So which side of the coin landed here?
Or is now Bad Good, Black White and hopefully you'll get run over at the next
Zebra crossing?

Thanks

Roy
--
gentoo-dev@lists.gentoo.org mailing list
 
Old 06-13-2008, 08:16 AM
Duncan
 
Default

Ciaran McCreesh <ciaran.mccreesh@googlemail.com> posted
20080613062612.46931b33@googlemail.com, excerpted below, on Fri, 13 Jun
2008 06:26:12 +0100:

> On Fri, 13 Jun 2008 09:30:54 +0530
> "Arun Raghavan" <arunisgod@gmail.com> wrote:
>> And why do you have to be plain insulting about it? Nobody can
>> magically spot every single bug in any piece of code presented to them.
>> In fact it's why the "given enough eyes ..." adage is one of the bases
>> of open source development.
>
> Which is why any responsible person ensures good test coverage.
>
>> I _honestly_ do not understand why there is so much trouble in simple
>> cooperation amongst adults.
>
> I agree entirely. Why the pkgcore people refuse to do basic automated
> tests is completely beyond me.

That may or may not be, but it's beside the point. The point is that a
bug was found, that fact was stated, and regardless of other points that
could be made, the developer of the code in question was all but forced
to call the person who caught the bug God and ask forgiveness for his
sin, in ordered to find out what the bug was.

Cooperation is understanding that people may have different development
methods and reporting the bug as found so it can be fixed, possibly
pointing out while doing so how much simpler it would be to find such
bugs in the future if an automated test case was created. Cooperation is
not forcing them to do it my way now, or at least admit my way's better,
before deigning to reveal the bug I know and they don't. If enough bugs
happen due to the lack of those tests and they hit enough people, the
problem will one way or another take care of itself as the test cases are
either provided and integrated somehow some way, or people move on to
more stable solutions. If not, perhaps those test cases weren't so vital
after all, and fixing the handful of bugs as they appeared ultimately
worked just as well as doing all those extra corner-case tests.

--
Duncan - List replies preferred. No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master." Richard Stallman

--
gentoo-dev@lists.gentoo.org mailing list
 
Old 06-13-2008, 08:26 AM
Ciaran McCreesh
 
Default

On Fri, 13 Jun 2008 08:16:57 +0000 (UTC)
Duncan <1i5t5.duncan@cox.net> wrote:
> > I agree entirely. Why the pkgcore people refuse to do basic
> > automated tests is completely beyond me.
>
> That may or may not be, but it's beside the point. The point is that
> a bug was found, that fact was stated, and regardless of other points
> that could be made, the developer of the code in question was all but
> forced to call the person who caught the bug God and ask forgiveness
> for his sin, in ordered to find out what the bug was.

No no, calling me God won't get anyone anywhere...

He was forced to do the kind of extremely basic testing that should
have been done before an EAPI 1 accepting package manager was put in
the tree. Unfortunately, he then committed a fix and didn't add unit
tests to prevent future regressions, nor did he add unit tests to cover
the rest of EAPI 1 functionality.

There is a big difference between obscure bugs and blatant
irresponsibility here.

--
Ciaran McCreesh
 
Old 06-13-2008, 09:01 AM
Patrick Lauer
 
Default

Ciaran McCreesh wrote:

On Fri, 13 Jun 2008 09:30:54 +0530
"Arun Raghavan" <arunisgod@gmail.com> wrote:


And why do you have to be plain insulting about it? Nobody can
magically spot every single bug in any piece of code presented to
them. In fact it's why the "given enough eyes ..." adage is one of the
bases of open source development.



Which is why any responsible person ensures good test coverage.


Just to pour some oil on the flames -

Y'all are aware that paludis can't parse a valid make.conf and does
ignore package.keywords at times, yes?


Test case is:

FEATURES="strict" # test and stricter fail

in make.conf ... if you had the tests you claim others lack that would
have been fixed a long time ago.

So please stop trolling when you fail so badly at it.


I _honestly_ do not understand why there is so much trouble in simple
cooperation amongst adults.



I agree entirely. Why the pkgcore people refuse to do basic automated
tests is completely beyond me.


Mirror, mirror on the wall.

--
gentoo-dev@lists.gentoo.org mailing list
 
Old 06-13-2008, 09:07 AM
"Fernando J. Pereda"
 
Default

On 13 Jun 2008, at 11:01, Patrick Lauer wrote:


Ciaran McCreesh wrote:

On Fri, 13 Jun 2008 09:30:54 +0530
"Arun Raghavan" <arunisgod@gmail.com> wrote:


And why do you have to be plain insulting about it? Nobody can
magically spot every single bug in any piece of code presented to
them. In fact it's why the "given enough eyes ..." adage is one of
the

bases of open source development.



Which is why any responsible person ensures good test coverage.


Just to pour some oil on the flames -


Then don't do it. You are doing a very bad marketing for the pkgcore
guys with your whinnings.



Y'all are aware that paludis can't parse a valid make.conf and does
ignore package.keywords at times, yes?


Test case is:

FEATURES="strict" # test and stricter fail

in make.conf ... if you had the tests you claim others lack that
would have been fixed a long time ago.


Yes, we are aware of that bug in a feature we consider highly
experimental.


You are amusing...

- ferdy

--
gentoo-dev@lists.gentoo.org mailing list
 
Old 06-13-2008, 09:08 AM
Ciaran McCreesh
 
Default

On Fri, 13 Jun 2008 11:01:19 +0200
Patrick Lauer <bugs@dev.gentooexperimental.org> wrote:
> Just to pour some oil on the flames -
>
> Y'all are aware that paludis can't parse a valid make.conf and does
> ignore package.keywords at times, yes?

Yep. We don't claim to or aim to completely support Portage configs.

> Test case is:
>
> FEATURES="strict" # test and stricter fail
>
> in make.conf ... if you had the tests you claim others lack that
> would have been fixed a long time ago.

No, we just don't bother supporting it. Remember that configs aren't
part of PMS.

Also note that if you had something like this in package.use:

foo/bar baz # monkey

Portage would until relatively recently (and after that section of PMS
was written, for the profiles side of it) set USE="baz # monkey".
Paludis chose to indicate an error rather than accept clearly nonsense
input.

There's no PMS violation here -- user configs aren't covered (and if
you do use a Portage user configuration with Paludis, you get a big fat
warning saying "this probably won't work, file tickets if you want
stuff fixed"), and PMS restricts profile use files to behaviour safely
supported by all EAPI 0 accepting Portage versions.

--
Ciaran McCreesh
 
Old 06-13-2008, 09:16 AM
Patrick Lauer
 
Default

Fernando J. Pereda wrote:



Just to pour some oil on the flames -


Then don't do it. You are doing a very bad marketing for the pkgcore
guys with your whinnings.

Dude. Shut up.

I'm not a pkgcore guy. If anything I'm a portage supporter. That I
accidentally host pkgcore.org doesn't mean I'm "one of them".



Y'all are aware that paludis can't parse a valid make.conf and does
ignore package.keywords at times, yes?


Test case is:

FEATURES="strict" # test and stricter fail

in make.conf ... if you had the tests you claim others lack that
would have been fixed a long time ago.


Yes, we are aware of that bug in a feature we consider highly
experimental.
Hmm, I'd have guessed config files are moderately relevant. And why
don't y'all fix a bug like that? First you insult others for not doing
tests, then you show a lack of tests and are proud of it. Augh.


You are amusing...

Hey, I gave you a testcase - now fix it, chop chop!

--
gentoo-dev@lists.gentoo.org mailing list
 
Old 06-13-2008, 09:22 AM
Ciaran McCreesh
 
Default

On Fri, 13 Jun 2008 11:16:31 +0200
Patrick Lauer <bugs@dev.gentooexperimental.org> wrote:
> > Yes, we are aware of that bug in a feature we consider highly
> > experimental.
>
> Hmm, I'd have guessed config files are moderately relevant.

You didn't notice the large warning telling you not to use Portage
config files?

> And why don't y'all fix a bug like that?

We do what PMS requires regarding handling of inline comments (which is
the same as what some EAPI 0 accepting Portage versions do, so PMS
can't allow inline comments), and indicate an error (rather than
writing junk, as older Portage did) when inline comments are used.

--
Ciaran McCreesh
 
Old 06-13-2008, 09:47 AM
"Fernando J. Pereda"
 
Default

On 13 Jun 2008, at 11:16, Patrick Lauer wrote:
Then don't do it. You are doing a very bad marketing for the
pkgcore guys with your whinnings.



I'm not a pkgcore guy. If anything I'm a portage supporter. That I
accidentally host pkgcore.org doesn't mean I'm "one of them".


Were you able to read English you'd have noted that I implicitly
excluded you from the "pkgcore guys" in that sentence.


Yes, we are aware of that bug in a feature we consider highly
experimental.
Hmm, I'd have guessed config files are moderately relevant. And why
don't y'all fix a bug like that? First you insult others for not
doing tests, then you show a lack of tests and are proud of it. Augh.


Use of Portage configuration files will lead to sub-optimal
performance and loss of functionality. Full support for Portage
configuration formats is not guaranteed; issues should be reported via
trac.


That's the pretty nice warning. Full support is not guaranteed. We do
take sane patches, however.


Stop flaming, please.

- ferdy

--
gentoo-dev@lists.gentoo.org mailing list
 
Old 06-13-2008, 09:53 AM
Patrick Lauer
 
Default

Ciaran McCreesh wrote:

On Fri, 13 Jun 2008 11:16:31 +0200
Patrick Lauer <bugs@dev.gentooexperimental.org> wrote:

Yes, we are aware of that bug in a feature we consider highly
experimental.


Hmm, I'd have guessed config files are moderately relevant.



You didn't notice the large warning telling you not to use Portage
config files?

I did. But how else can I compare things or move back to portage if I
don't like it?



And why don't y'all fix a bug like that?



We do what PMS requires regarding handling of inline comments (which is
the same as what some EAPI 0 accepting Portage versions do, so PMS
can't allow inline comments), and indicate an error (rather than
writing junk, as older Portage did) when inline comments are used.

So you say the thing you wrote excludes things you don't like so you can
claim things by referencing it as authoritative.


Does anyone else think that maybe there's a slight conflict of interest
there?


I hope that PMS, as it stands now, does not become a standard. It is
obviously very leaky and ignores issues so that you can claim PMS
compatibility without being compatible to each other.


--
gentoo-dev@lists.gentoo.org mailing list
 

Thread Tools




All times are GMT. The time now is 04:04 PM.

VBulletin, Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Content Relevant URLs by vBSEO ©2007, Crawlability, Inc.
Copyright 2007 - 2008, www.linux-archive.org